Consumer purchases of energy-efficient cars: behavioural implications for policy

Transcription

1 Consumer purchases of energy-efficient cars: behavioural implications for policy Ibon Galarraga, Josu Lucas and Steffen Kallbekken XII CONGRESO DE LA ASOCIACIÓN ESPAÑOLA PARA LA ECONOMÍA ENERGÉTICA Salamanca, 2017

2 Overview 1. Introduction 2. Problem: Energy Efficiency Gap 3. Existing evidence: Car market in Spain 4. Concluding remarks 2

3 1. Introduction 2. Problem: Energy Efficiency Gap 3. Existing evidence: Car market in Spain 4. Concluding remarks 3

4 The importance of energy efficiency Efforts to reduce fossil-fuel use in the different sectors (e.g. transport, industry, building). European Commission identifies increased energy-efficiency (EE) as the most cost-effective and rapid way to reduce CO2 emissions. The IEA estimates that EE measures can reduce global CO2 emissions by up to 10 15% per year at no direct additional cost. IPCC (2014) suggests an investment in energy efficiency (EE) in transport, industry and building of 336 billion US$. EU Climate and Energy package that sets the target of reducing energy consumption by 20% by

5 The importance of energy efficiency Transport sector accounts for 14.3% of global GHG emissions, Building for almost 20% of global GHG emissions and industry 31%. The EU goal of a 27% energy saving in the residential sector (European Council 2006). 5

6 1. Introduction 2. Problem: Energy Efficiency Gap 3. Existing evidence: Car market in Spain 4. Concluding remarks 6

7 Some problems: BUT, Energy Efficiency Paradox exists (Howarth and Andersson, 1993; Jaffe et al., 2009): Private investments in energy efficiency that seem to be economically worthwhile are not always made. And, Some individuals make investments in EE when economically they would not appear to be worthwhile. Can be explained: insufficient information, principal-agent problems, Cannot know efficiency, hidden costs Owner/tenant lack of access to capital, divergences between social and private discount rates, consumer behaviour that is motivated by non economic factors, such as a desire to contribute to a public good. 7

8 Energy efficiency paradox: (Ramos et al., 2015) 8

9 Instruments: Labels and certification (Ramos et al., 2015) Labels might help overcoming information failures: Incomplete and/or asymmetric. Transaction costs. Uncertainties. Also behavioural failures: Limited attention. Aversion to uncertainty. And finally, principal agent problems. Great potential (Ramos et al., 2015). More effective to show energy savings or economic losses than potential benefits. 9

10 1. Introduction 2. Problem: Energy Efficiency Gap 3. Existing evidence: Car market in Spain 4. Concluding remarks 10

11 EU policy in Spain: The EU has energy labels for domestic appliances since In 1999 this was extended to include cars (Directive 1999/94/CE). Retailers have to display characteristics of the car (size, consumption and CO2 emissions). Also a voluntary comparative labelling scheme with different categories of energy efficiency (from A, the most efficient, to G, the least efficient). The label also can include other information (running costs, annual tax costs, or additional attributes of the vehicle). Thus, major differences between labels in different countries. In Spain the Directive was transposed by Royal Decree 837/2002, and today all car retailers have to show both the standard EU label and the comparative label for their vehicles. 11

12 EU policy in Spain: In EU countries 40% disagreed with the statement that it was easily recognisable; and 44.5 % agreed that car labels were unfamiliar to them. The differences also include the way in which categories of efficiency are calculated. (Codagnone et al, 2013) Absolute labelling scheme for all the cars in the market: the most efficient cars which pollute the least, usually the smallest cars, are labelled A class, while other cars, bigger or less efficient, are labelled B, C, D, E or G. In most European countries, including France, Belgium, Denmark and the United Kingdom (Brannigan et al, 2011). Relative labelling (Brannigan et al, 2011) the label of the car depends on how much the fuel consumption and emissions of the car deviates from the average within its market segment (for instance small, mini, small sedan, big sedan, etc.). Spain and Germany. 12

13 EU policy in Spain: PIVE The PIVE (Programa de Incentivos al Vehículo Eficiente) plan was implemented in 2012 and is currently in its 7th edition in The PIVE subsidises the purchase of cars categorised as class A or B, electric cars, and cars which use gas or other alternative fuels. The subsidy is only applicable to cars up to a maximum price of 25,000, and it consists of a minimum discount of 1,000 in the price before taxes, which the producer or retailer has to apply, plus a subsidy of at least 1,000 after taxes financed by public funding earmarked for the PIVE. To the best of our knowledge there are, as yet, no studies assessing the impact of the policy. 13

14 Energy efficiency in Spanish car market: To the best of our knowledge, there are no official statistics on the energy efficiency class of the new light-duty vehicles sold in the Spanish market. The National Association of Car and Truck Producers (Asociación Nacional de Fabricantes de Automóviles y Camiones, ANFAC) offers monthly data on the number of cars sold, but does not collect information on the energy efficiency performance of the cars sold. As a supplement to this information, The Spanish Energy Diversification and Saving Institute (Instituto de Diversificación y Ahorro Energético, IDAE) offers a list of the cars and models available and their energy efficiency attributes. 14

15 41% of the cars in 2012 were very efficient (A class). Table 1: Number of cars sold in Spain in 2012 per market segment, and their energy efficiency A and B label cars make up more than 75% of all cars sold Most of the cars sold were small (27.8%) or small sedans (27.3%). n. cars % % A class % B class % Others Unknown Small 194,616 27,82% 37,68% 50,70 % 11,62% 1,05% Mini 35,164 5,03% 25,16% 38,39 % 36,45% 0,58% Small 26,11 % Sedan 191,604 27,39% 53,40% 20,49% 0,13% Big Sedan 85,310 12,19% 69,95% 18,75 % 11,30% 0,05% Small 44,16 % Minivan 75,565 10,80% 42,51% 13,33% 0,58% Big class A. The energy efficiency of SUVs32,16 was% very low. Minivan 10,573 1,51% 8,67% 59,17% 3,51% Sport 2,176 0,31% 1,30% 21,61 % 77,09% 19,90% Luxury 1,581 0,23% 52,16% 40,68 % 7,16% 33,08% (and previous) PIVE rebate schemes? Executive 10,806 1,54% 33,98% 46,33 % 19,69% 26,37% Small SUV 30,177 4,31% 2,97% 21,90 % 75,13% 2,64% Medium influenced the 18,72 % SUV 52,198 high proportion 7,46% 5,30% of efficient cars 75,98% sold? 1,25% Big SUV 2,757 0,39% 0,00% 0,00 % 100,00% 0,40% Luxury 31,00 % SUV 7,062 1,01% 0,00% 69,00% 29,51% TOTAL 699,589 41,07% 34,20 % 24,73% 1,53% The share of sport and luxury cars was very low. The proportion of efficient cars varies from one segment to another: the proportion of Sport, all types of SUV and Big Minivans with class A was very low, while more than half of all small and big sedans were Are the high sales of efficient cars a consequence of the current Perhaps other factors such as high fuel prices might also have How frequently the labels granted in a labelling scheme are reviewed also has an important effect on the proportion of efficient cars sold. In Spain annually. Source: Own calculations using data from IDAE and ANFAC. 15

16 Energy Efficiency private vehicles - Spain Study: Galarraga et al. (2014) cars Cross-sectional data with more than observations containing official prices and a set of detailed vehicles characteristics, including the energy efficiency label. Subsample of almost 400 observations with retail prices (gathered by the Mystery Shopping method) performed by an specialized survey company during September-November Each observation was matched with its correspondent EE label from the IDAE database. Methodology: Hedonic price approach with mystery shopping. Results: A statistically significant coefficient of the variable that measures the effect of (A, B) energy-efficiency labels: 3%-5.9% price premium (official listing and mystery shopping ). 16

17 Energy Efficiency private vehicles - Spain Compare WTP for a labelled A vehicle during the 10 years expected lifetime with the present value of the corresponding energy savings. WTP for and savings from energy-efficient vehicles Discounted fuel savings WTP for a vehicle labelled A, using the average price for the official-price subsample WTP for a vehicle labelled A, using the average price for the retail-price sample r= 5% Euros r= 10% Euros ,6 r=15% Euros Consumers undervaluing EE? Energy efficiency paradox? Overestimation of WTP? 17

18 Energy Efficiency private vehicles Spain We consider two alternative hypotheses about how consumers make this decision. This is a simplification as many consumers make decisions using simultaneous or nested processes (Noblet et al., 2006). But this simplification does fit well with the policy analysis in Brannigan et al., (2011) and serves well to explore the implications of choosing one type of labelling or the other. As far as we are aware no empirical studies are available to support the type of labelling chosen in EU Member States. If such studies existed they could have offered some insights on how purchasing decisions are made in each country. 18

19 Energy Efficiency private vehicles Spain Absolute decision: Concerned consumers select the most energy efficient car in the market independent of segment. Relative decision: Consumers first decide what type of car (i.e. the segment) and then choose the efficiency within the segment. A third way, consumers who select the brand and even the model, and then within those options select the most efficient one. Not discussed here. Let us estimate (calibrate) elasticities! Quantity Based Demand System! (Galarraga and Markandya, 2004) 19

20 Quantity Based Demand System V i : demand for quality i (energy efficiency) of good V (appliance) in comparable units. That is: P i : price of quality i of good V. M: total expenditure. P: aggregate price of good V w j : expenditure share of good V. V i V = β i( P i P ) The demand for quality i of good V can be defined as Where β i 0 is a constant and α 0 is the price sensitivity If we now define a price index P as P = s P i i i where s i 0 and s i = 1 ii = α + α μ s i ij = α μ s j 20

21 Quantity Based Demand System We need own elasticity of less efficient (other) cars, the income elasticity of demand for cars and the expenditure shares for both efficient and non-efficient (or less efficient) cars. Own price elasticities from to -1.2 and income elasticity 0.3, 0.5 and 1. (Whelan, G.A., 2007, Hymans, 1971 and Matas and Raymond, 2008). Expenditure shares for non-efficient cars from Spain s National Office of Statistics. We use the price premium estimate of % of the average car price (Galarraga et al., 2014) to calculate the expenditure shares for efficient cars (class A). The expenditure shares are: WO= WA= WX=

22 Absolute : Income least 1, 0.5 and 0.3 Table 2C: Own and cross price elasticities of demand QBDS (Income elasticity = 0.3) Own O Cross OA Own A Cross AO The demand for efficient cars (A) is slightly more elastic than demand for non-efficient cars (O). The demand for efficient cars decreases (increases) more than demand for non-efficient ones when the price of cars increases (decreases). The cross effects also suggest that changes in the demand for efficient cars are greater than the effect on other, less efficient ones. This difference increases as the price elasticities increase. 22

23 Table 4c: Own and cross price elasticities of demand per segment (Income elasticity = 0.3) SEDAN SPORT & LUXURY Own O Cross OA Own A Cross AO Own O Cross OA Own A Cross AO MINI SMALL Own O Cross OA Own A Cross AO Own O Cross OA Own A Cross AO MINIVAN SUV Own O Cross OA Own A Cross AO Own O Cross OA Own A Cross AO Relative: Income least 1, 0.5 and 0.3 The demand for the most efficient cars (class A) is less elastic than demand for non-efficient cars. This result is driven by the fact that the proportion of efficient vehicles in the market is lower than that of non-efficient ones for all segments except sedans. The range of elasticity values varies significantly for the cases of Mini, Sports and Luxury and SUV vehicles 23

24 1. Introduction 2. Problem: Energy Efficiency Gap 3. Existing evidence 4. Concluding remarks 24

25 Concluding remarks Energy efficiency is part of the long term climate solution, and it is smart way if saving resources. Labels, audits, feedback, taxes, subsidies, standards and many other instruments exist. We need to combine them well! Energy labelling is one of my favourites. And is acquiring a major importance in the light of the EU Climate and Energy. It can be used to reduce information asymmetries but also to support other policy instruments such as taxes and subsidies. Many examples exist in EU. Policies should be well designed and it is not always the case. 25

26 Concluding remarks With Absolute decision the demand for vehicles with higher efficiency level are greater than that for less efficient ones. Pricing policies are likely to be more effective when applied to A labelled cars and therefore policies based on pricing systems may have a role to play in incentivising the purchase of more efficient vehicles. When relative decision is assumed, that is, when consumers choose the car segment first and then the energy performance, the opposite result is found. Additionally, in all but one of the cases the cross price elasticities AO are greater than cross OA, which means that impacts of changes in prices of the labelled car segment affect the demand for less efficient ones much less than in the opposite direction. This is an expected result when the share of nonefficient vehicles is greater than the share of A labelled ones. 26

27 Concluding remarks The exception to this is the case of A labelled Sedan vehicles with a greater share in this market segment that makes the cross elasticticity AO lower than the cross OA. That is, in this case changes in prices in A labelled cars affects the demand of nonlabelled ones more. This effect cannot be noticed when showing values under absolute decision making hypothesis because the impacts on the rest of the segments overturn this. 27

28 BC3 Contact: Thank you! 28